Discussion
Although the trade-off between within-host infection load and transmission is a central tenet of pathogen evolution (Alizon et al., 2009), remarkably little is understood of this association in natural populations, and under coinfection scenarios that are prevalent across pathosystems (Tollenaere et al., 2016, Alizon et al., 2013). Here, we study within-host disease load and between-host transmission experimentally to understand how sensitive their association is to pathogen strain identity and coinfection with a virus. We conduct a survey of infection across 260 wild host populations to test whether our experimental results are reflected in epidemiological patterns in the wild.
Phomopsis subordinaria was detected in nearly half of the surveyed natural host populations. The other fungal pathogen studied in this same host population network - the powdery mildew fungus P. plantaginis - infects annually 2-20% of P. lanceolatapopulations (Jousimo et al., 2014), and hence, by comparison P subordinaria is relatively common. In few other wild plant pathosystems, similar disease incidence rates have been observed (e.g.Triphragmium ulmariae rust infected 29-69% Filipendula ulmaria host populations (Zhan et al., 2018), and Uromyces valerianae rust infected 43-73% of Valeriana salina populations (Ericson et al., 1999)). In our study, spatial structure was the main determinant for pathogen incidence as both host population connectivity and regional district explained variation in P. subordinariadistribution. Positive correlation between host connectivity and infection incidence suggests that host population connectivity increases between population transmission, as predicted by metapopulation theory (Hanski, 1999). Large populations were more likely to be infected than small populations, which is also in line with theoretical predictions (Hanski, 1999). Unlike in P. plantaginis (Jousimo et al., 2014), host population size did not have an effect on within-population infection prevalence. While airborne pathogens such as P. plantaginis are expected to be sensitive to host population size through density-dependent transmission, P. subordinaria is vector-transmitted and hence, is expected to be less responsive to variation in host population size (Thrall et al., 1995). This result highlights the importantance of the mode of transmission is for epidemics.
Our field survey releveled extremely high within-host infection load with 84% of host inflorescences being infected on average. While within-host infection load was at its upper limits, transmission was not, as the proportion of uninfected hosts was large. Majority of epidemics were limited to less than 10% of hosts being within populations. We did not find evidence of a trade-off between within-host infection load and transmission limiting the spread of P. subordinaria within its host populations. Instead, we found a positive correlation between within-host disease load and population infection prevalence. The observed low within-host populations prevalence ofP. subordinaria could also result from a range of other factors unrelated to life-history correlations, including variation in host plant susceptibility or pathogen infectivity (de Nooij & Damme, 1988), constrained vector transmission (de Nooij, 1988, Pleydell et al., 2018), and interactions with other pathogens than PlLV (Susi et al., 2015a, Susi et al., 2015b).
Cross-kingdom coinfections are common (Lawn et al., 2006, Chen et al., 2020) (Tollenaere et al., 2016, Tollenaere et al., 2017, Telfer et al., 2010), and they are often suggested to have serious consequences in disease epidemics and disease severity. The trade-offs observed under single host–single pathogen scenarios may change under coinfection, as host exploitation rates are expected to change under diverse infections (Alizon et al., 2009). Plantago lanceolata is a host for a number of pathogens in the Åland Islands, and coinfections are frequently observed (Susi et al., 2015a, Susi et al., 2019). To understand how coinfections may shape life-history correlations and disease dynamics ofP. subordinaria , we studied disease development under coinfection with a recently characterized virus, PlLV. We found that coinfection with virus had a profound impact on within-host infection load and transmission of P. subordinaria that can further impact evolution and epidemiology of the pathogen. Coinfection alleviated the harm caused for the host and increased transmission potential. Under natural epidemics such trade-offs could translate into low within-host disease load and increased among host transmission.
The two strains differed significantly in their within-host infection load, with strain P43 outperforming strain P29. Significant variation among pathogens strains in their life-history traits is commonly observed in natural pathogen populations (Tack et al., 2012). For both strains, within-host infection load was lower under coinfection with PlLV than when P. subordinaria infected the host alone, while transmission potential was significantly higher under coinfection than under single infection. The relationship between within-host infection load and transmission potential was mediated by both strain identity and coinfection. We observed a negative correlation between the measured life-history stages in strain P29 suggesting a trade-off, whereas in strain P43 there was no evidence of a trade-off. This result is in line with previous research that found life-history correlations to be depending on the pathogen genotype (Clement et al., 2012, Bruns et al., 2014). The negative relationship between within-host infection load and pycnidia formation became positive under coinfection. This is in line with an earlier study testing coinfection with two strains of powdery mildew fungus P. plantaginis where the strains had higher performance and positive life-history trait correlations when challenged with a competing strain (Laine & Mäkinen, 2018). Contrary to other studies on coinfection where response to coinfection has been found pathogen strain specific, here the response of the P. subordinaria strains was similar. In bacteria –fluke coinfections on salmons (Louhi et al., 2015), genotype specific responses on coinfection were observed. Similarly, virus strain combinations resulted in different within-host growth rates of bacterial pathogen on rice (Tollenaere et al., 2017).
This study increases our understanding on the factors generating diversity in epidemics in natural populations. Furthermore, this is one of the very first reports addressing the knowledge gap on how pathogen life-history traits correlate in realized epidemics. By showing that pathogen coinfection and strain identity may alter life history correlations this study contributes to better understanding of disease evolution and epidemiology.